Soy isoflavone trials for menopausal symptoms have produced famously inconsistent results. For years that inconsistency was attributed to study design.
There’s a better explanation, and it isn’t about the trials. It’s about who was in them.
The conversion step
Daidzein is one of the two main soy isoflavones. In some people, gut bacteria convert it further.
Equol is a terminal metabolite of daidzein, and it has higher affinity for estrogen receptor beta than daidzein itself, exhibiting stronger estrogen-like activity[1].
The conversion runs through intermediates — dihydrodaidzein, then tetrahydrodaidzein — each requiring specific bacterial enzymes[1]. This capability classifies individuals as equol producers or non-producers[1].
Two facts make this consequential.
Equol is not found in food[2]. It exists only if your gut makes it.
And equol is more stable than other flavonoid metabolites, existing in a higher proportion as free form in the body — meaning higher blood concentrations sustained longer[3].
So the same soy intake produces materially different internal exposure depending on which bacteria you host.
The prevalence gap
This is where it becomes a population-level issue.
| Population | Producers |
| Western adults | 20–35% |
| East Asian adults | 50–60% |
| Korean study (Akaza) | 59% |
| Korean GWAS cohort (n=1,391) | 70.1% |
[Table 1] Equol producer prevalence by population · Source: Synthesis of related studies[1][4][5][6]
The Korean figure deserves emphasis. In a study measuring serum equol in 1,391 Korean subjects, 70.1% produced equol[4].
That’s roughly two to three times the Western rate.
The Korean study also found something worth noting clinically: blood pressure was significantly lower in equol producers (β ± SE = −1.35 ± 0.67, p = 0.045)[4].
Diet, not genes
An important mechanistic point, because it changes what the difference means.
The difference in equol producer rates between Asian and Western countries is believed to be due to the gut microbiome, not genetics[7].
The dietary contrast is stark. Daily soy isoflavone intake is 25–50 mg/day in Japan and other East Asian countries versus about 2 mg/day in Western countries[5].
And the relationship appears bidirectional. A Japanese study of 1,044 adults found equol producers were significantly older than non-producers, with both producer prevalence and daily daidzein intake increasing with age[8].
The bacterial correlates were measured directly. Among producers, A. celatus was present in 50.2% and S. isoflavoniconvertens in 38.9%, versus 11.8% and 4.3% in non-producers[8].
A caveat on stability. Producer status is not necessarily permanent — recent work has examined its stability in relation to gut microbiota diversity[5], which implies it can change. Whether it can be reliably induced is a separate and unsettled question.
What this does to trials
Here’s the methodological consequence, and it’s the reason this topic belongs in a practitioner post.
A randomised trial of soy isoflavones in a Western population is randomising a group in which roughly 70–80% cannot produce the active metabolite.
If the effect is concentrated in producers, that trial is diluting a real signal across a majority who cannot respond. The result looks like a small or absent effect.
The cognition literature demonstrates this directly. Epidemiological studies in East Asia show dietary isoflavone intake inversely associated with cognitive decline, while randomised trials in Western countries did not generally show cognitive benefit[7].
The authors attribute the discrepancy to producer capability: 40–70% of East Asians produce S-equol, whereas 20–30% of Westerners do[7].
And a subgroup analysis tested it. In the WISH trial, S-equol producers showed improved cognition compared to the control group[7]. A Japanese cross-sectional study found S-equol production significantly inversely associated with mild cognitive impairment[7].
This is a specific instance of a general problem covered throughout this series: a null result in a population that cannot respond is not evidence that the intervention doesn’t work — it’s evidence about who was studied.
A bit more detail — on why this pattern is hard to escape. Randomisation distributes producer status evenly between arms, which is methodologically correct but doesn’t solve the dilution. The fix is stratification by producer status at enrolment, or pre-specified subgroup analysis. Subgroup findings carry the usual caveats — as covered in the vitamin D post, an unreplicated subgroup result is weak evidence. But when the subgroup is defined by a measurable metabolic capacity with a known mechanism, that’s a stronger basis than a post-hoc split.
Testing and the direct-supplement route
Two practical developments follow.
Producer status is measurable. The Korean GWAS determined it by serum equol concentration measured by LC-MS/MS[4]; other studies use 24-hour urinary equol excretion after a soy challenge[9]. These are research methods, and availability outside research settings varies.
S-equol can be supplied directly. Non-producers who want equol’s effects can take S-equol supplements, bypassing the need for gut conversion[2].
That’s a logically clean solution. It also relocates the question rather than settling it. A directly-supplied S-equol product falls under whatever regulatory framework applies where it’s sold. And as covered in the claims comparison post, Korea’s recognised claim for soy isoflavone covers bone and joint health rather than menopausal symptoms.
For readers in Korea
The prevalence data cuts in a specific direction here.
A Korean population is more likely than a Western one to include equol producers — 59% in one study, 70.1% in another. If you’re reading Western trial results and finding them discouraging, those results were generated in a population where most participants could not produce the metabolite.
That doesn’t mean soy isoflavone works for you. It means the most commonly cited evidence was generated under conditions that don’t describe you.
And as covered in the dietary soy post, habitual isoflavone intake here is already substantial — a usual intake of around 47 mg/day, with peak intake in the 50–64 age band. The population that produces equol is also the population already consuming the substrate.
Closing
The reason this topic matters beyond soy is what it demonstrates about reading trial evidence.
For years, mixed results on soy isoflavones were attributed to methodological weakness. The alternative explanation is that the trials were fine and the populations were wrong — that roughly three-quarters of Western participants lacked the bacterial capacity to produce the active compound.
If that’s right, then “the evidence is mixed” was never the correct summary. The correct summary is that the effect requires a metabolic step most trial participants couldn’t perform.
Which is a different kind of uncertainty, and one with an actual resolution: measure producer status, or supply the metabolite directly.
It also raises an uncomfortable general question. How many other ingredients have a conversion step nobody has identified yet — where the “inconsistent evidence” is really unrecognised population heterogeneity?
Key Terms
- Daidzein — one of the two principal soy isoflavones.
- Equol — a gut bacterial metabolite of daidzein with stronger estrogen receptor beta activity than daidzein itself.
- Equol producer — a person whose gut microbiota can perform the conversion.
- Estrogen receptor beta (ERβ) — one of two estrogen receptor types; the one equol preferentially binds.
- Stratification — dividing trial participants by a characteristic before randomising.
- Population heterogeneity — systematic differences within a study population that affect response.
At a Glance
- Equol is not found in food — it exists only if gut bacteria convert daidzein
- It has higher affinity for estrogen receptor beta than daidzein and is more stable in the body
- Prevalence: 20–35% of Western adults, 50–60% of East Asians
- Korean data: 59% in one study, 1% in a GWAS cohort of 1,391
- The difference is attributed to gut microbiome, not genetics — soy intake is 25–50 mg/day in East Asia versus ~2 mg/day in the West
- Producers carried *A. celatus* (50.2%) and *S. isoflavoniconvertens* (38.9%) versus 8% and 4.3% in non-producers
- East Asian epidemiology shows cognitive benefit while Western RCTs generally did not — attributed to producer capability
- S-equol can be supplied directly, bypassing the conversion requirement
※ This article discusses metabolic variation and trial interpretation for general information only. It does not replace medical advice, and it is not a recommendation to take soy isoflavone or S-equol products. Isoflavones have oestrogenic properties and may interact with thyroid function — if you have a hormone-sensitive condition, thyroid disease, or take related medication, consult a clinician before use. Approved claims differ by country; confirm current wording on official sources.
References
- “A cross-sectional study of the gut microbiota associated with urinary and serum equol production status in a general population of Japanese men”, ScienceDirect (conversion pathway; ERβ affinity; 50–60% versus 20–30% prevalence; dietary intake comparison), https://www.sciencedirect.com/science/article/pii/S2212429225012246
- “Equol for Perimenopause: The Soy Isoflavone Metabolite” (equol absent from food; direct S-equol supplementation route), Happy Aging, https://happyaging.com/blogs/news/equol-soy-isoflavone-metabolite-perimenopause-2026
- “Equol: a metabolite of gut microbiota with potential antitumor effects”, Gut Pathogens (stability and free-form proportion; global prevalence range), https://link.springer.com/article/10.1186/s13099-024-00625-9
- “Epidemiological profiles between equol producers and nonproducers: a genomewide association study of the equol-producing phenotype”, PMC (Korean cohort n=1,391, 70.1% producers, blood pressure finding, Akaza 59% figure), https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3448036/
- “Stability of equol production capability is associated with the diversity of the gut microbiota of the host: a prospective cohort study”, BMC Microbiology (prevalence figures; dietary intake comparison; stability question), https://link.springer.com/article/10.1186/s12866-026-04749-7
- “Prevalence of the Equol-Producer Phenotype and Its Relationship with Dietary Isoflavone and Serum Lipids in Healthy Chinese Adults”, PMC (Western 20–35% versus Asian 50–55%), https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3900832/
- “Potential Protective Mechanisms of S-equol on Cognitive Decline and Dementia”, PMC (epidemiology versus RCT discrepancy; microbiome not genetics; WISH subgroup analysis), https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9570153/
- “Daidzein Intake Is Associated with Equol Producing Status through an Increase in the Intestinal Bacteria Responsible for Equol Production”, Nutrients (1,044 subjects; bacterial prevalence figures; age association), https://www.mdpi.com/2072-6643/11/2/433
- “S-equol producing bacteria: isolation and identification from gut microbiota”, PMC (measurement approaches and conversion variability), https://pmc.ncbi.nlm.nih.gov/articles/PMC12953384/